In the United States (US), the rates of end-stage renal disease (ESRD), hospitalization and 30-day hospital readmission differ markedly by age, race-ethnicity, socioeconomic status and site of care. Despite the disproportionate burden of ESRD observed among persons of racial-ethnic minority groups and low-income populations, few US-based studies in chronic kidney disease (CKD) have successfully examined the health processes and outcomes of persons receiving care within public hospitals and safety-net health systems. The primary aims of this proposal are (1) to develop and validate an ESRD risk prediction instrument and compare its performance to existing ESRD risk prediction tools using data from persons with moderate to advanced CKD who received ambulatory care within public health systems; and (2) to estimate rates of, and identify factors associated with, 30-day hospital readmission among persons with moderate to advanced CKD within a large public hospital system. This proposal will utilize a retrospective cohort study design to develop risk prediction tools to enable health providers and health systems to systematically identify persons at highest risk for CKD-related morbidity. Such high-risk patients could receive more intensive surveillance, intervention, and earlier nephrology and transitional care. The long-term objective of this research is to reduce the burden of CKD among traditionally underserved populations using effective interventions that can be used to slow progression and prevent complications of CKD.